Effects of N:P:Si Ratios and Zooplankton Grazing on Phytoplankton Communities in the Northern Adriatic Sea

AQUATIC MICROBIAL ECOLOGY Vol. 18: 37-54, 1999 1 Published July 16 Aquat Microb Ecol

Effects of N:P:Si ratios and zooplankton grazing on phytoplankton communities in the northern Adriatic Sea. I. Nutrients, phytoplankton biomass, and polysaccharide production

Edna Granelil.*, Per ~arlsson',Jefferson T. urne er^,Patricia A. ester^, Christian Bechemin4, Rodger ~awson',Enzo ~unari'

'University of Kalmar, Department of Marine Sciences. POB 905, S-391 29 Kalmar. Sweden 'Biology Department. University of Massachusetts Dartmouth. North Dartmouth, Massachusetts 02747. USA 3National Marine Fisheries Service, NOAA, Southeast Fisheries Science Center. Beaufort Laboratory, Beaufort. North Carolina 28516, USA 4CREMA-L'Houmeau (CNRS-IFREMER), BP5, F-17137 L'Houmeau, France 'Chesapeake Biological Laboratory, University of Maryland, Solomons, Maryland 20688. USA '~aboratorio di Igiene Ambientale, Istituto di Sanita, Viale R. Elena 299. 1-00161 Rome, Italy

ABSTRACT: The northern Adriatic Sea has been historically subjected to phosphorus and nitrogen loading. Recent signs of increasing eutrophication include oxygen def~ciencyin the bottom waters and large-scale formation of gelatinous macroaggregates. The reason for the formation of these macroaggregates is unclear, but excess production of phytoplankton polysacchandes is suspected. In order to study the effect of different nutrient (nitrogen~phosphorus:silicon)ratios on phytoplankton production, biomass, polysacchandes, and species succession, 4 land-based enclosure experiments were performed with northern Adriatic seawater. During 2 of these experiments the importance of zooplankton grazlng as a phytoplankton loss factor was also investigated. Primary productivity in the northern Adri- atic Sea is thought to be phosphorus limited, and our experiments confirmed that even low daily phosphorus additions Increased phytoplankton biomass. However, this only occurred when nitrogen additions were high. Alternatively, when nitrogen was added in low concentrations, ulth simultaneous high phosphorus add~t~ons,phytoplankton biomass declined Nitrogen deficiency induced the highest production of polysacchandes per unit of cell carbon, while nutrient-sufficient and phosphorus-deficient treatments caused a higher production of polysaccharides in total. In order to decrease the frequency of algal blooms and high polysaccharide production in the northern Adnatic, it appears necessary to reduce the amounts of incoming nutrients. Since phosphorus has a high turnover rate in low P: high N waters of the northern Adnatic, and slnce our experiments show that a shortage of nitrogen can produce reduced levels of phytoplankton biomass and total polysaccharides, a reduction of the nitrogen discharge would probably be the best countermeasure for eutrophicat~onIn the northern Adnatic Sea.

KEYWORDS: Adriat~cSea. Nitrogen. Sdicon . Lunitation . Phytoplankton . Phosphorus. Polysaccharide Chemical composition

INTRODUCTION Adriatic Sea has been undergoing eutrophication for decades due to large loadings of phosphorus (P) and Eutrophication due to anthropogenic nutrient load- nitrogen (N) originating from agriculture, industries ing in coastal marine waters is a growing problem and sewage (Justic 1987). Although anoxic events in throughout the world (Likens 1972, Caraco et al. 1990, the bottom waters of the northern Adriatic Sea have Nixon 1990, Vollenweider et al. 1992). The northern occurred sporadically for several centuries (Piccinetti & Manfrin 1969), the frequency of occurrence and size of 'E-mail: [email protected] se the areas exhibiting oxygen deficiency have dramati-

O Inter-Research 1999 38 Aquat Microb Ecol 18: 37-54, 1999

Trieste For most marine coastal waters and inland seas N is thought to be the most limiting nutrient for phytoplankton production (Ryther & Dunstan 1971, Howarth 1988, Graneli et al. 1990, Oviatt et al. 1995). However, a few cases have been reported where P limits production (Berland et al. 1980, Smith 1984), and P has been suggested to be limiting in marine coastal waters only when large nutrient loads with high N:P ratios reach \ PO river '~'% coastal waters (Howarth 1988). Phosphorus has been suggested to be the most limiting nutrient for phytoplankton primary production in the northern Adriatic Sea. Indeed, 80% of the inorganic N:P ratios found in the surface waters suggest

-Sampling that P is limiting phytoplankton growth (Chiaudani et al. 1980b), and bioassay experiments have confirmed a '9 stat ion marked stimulation of phytoplankton growth when P was added (Pojed & Kveder 1977). The main objectives of our experiments were to Fig. 1. Location of the sampling station for the water used in investigate: (1) the effects of different N, P and silicon the enclosure experiments (Si) ratios and concentrations on phytoplankton growth, biomass, production, polysaccharide produc- cally increased during the last 2 decades (Justic 1987, tion, biochemical composition and changes in the 1991). intracellular content of carbon (C), N and P; (2) the Another possible sign of increasing eutrophication in effects of these nutrient manipulations on the abun- these waters is formation of gelatinous macroaggre- dance and composition of natural phytoplankton com- gates of huge proportions, known locally as 'mare munities; and (3) the impact of zooplankton grazing on sporco' or 'dirty water' (Stachowitsch et al. 1990, Mar- these phytoplankton communities. chetti 1992). These gelatinous masses are thought to Experiments were performed in May 1993, June be related to phytoplankton blooms, and have caused 1993, May 1994 and August 1994 in land-based enclo- beach fouling throughout the northern Adriatic Sea, sures in Fano, Italy. Nutrient, phytoplankton biomass, with substantial economical losses for an otherwise primary production and polysaccharide production flourishing tourism industry. These phenomena have results for the June 1993, May 1994 and August 1994 not been associated with severe oxygen deficiency, experiments will be presented in this paper. Effects of however, probably due to the low content of organic different nutrient ratios and concentrations on phyto- matter in the gelatinous mass (Marchetti 1992). How- plankton species composition in the June 1993 and ever, oxygen deficiency has been observed in the bot- May 1994 experiments will be presented in the second tom waters after large accumulations of gelatinous paper in this series (Carlsson & Graneli 1999, in this macroaggregates during the summers of 1988 and issue). Fluctuations of zooplankton in enclosures and 1989 (Degobbis 1989). Other harmful blooms occurring results of zooplankton grazing experiments from the in this area include those of toxic phytoplankton spe- May 1993 and June 1993 experiments will be pre- cies (Boni et al. 1993, Fonda Umani et al. 1993). sented in the third paper in this series (Turner et al. The main source of nutrients for the northern Adri- 1999, in this issue). atic Sea is the PO River (Degobbis & Gilmartin 1990, Vollenweider et al. 1992), with a mean discharge of about 1500 m3 S-' (Cati 1981) of water containing high MATERIALS AND METHODS concentrations of N and P. The total amount of P reach- ing the northern Adriatic is approximately 30000 tons Experiments were performed in land-based enclo- yr-l, of which the PO River contributes about 60% (Chi- sures using northern Adriatic seawater. The construc- audani & Vighi 1982). Approximately 152000 tons yr-' tion of this land-based enclosure system has been pre- of N reaches coastal waters, of which 75% is dis- viously described in Olsson et al. (1992). The water charged by the PO River (Chiaudani et al. 1980a). The containing the natural phytoplankton communities large transport of nutrients to the relatively shallow was pumped from 1 to 2 m depth from a station 15 km sea, markedly increases the primary production in off the east coast of Italy, offshore from Fano (Fig. l), these waters (Gilmartin & Revelante 1980, Malej et al. and filled into 500 1 dark-blue plastic containers. 1995). While filling the tanks on board the ship, the water Graneli et al.: Effects of nutrients and zooplankton grazing on phytoplankton. I 3 9

was filtered through 100 pm mesh in order to remove from the different 500 1 plastic transport containers extraneous mesozooplankton. We realized that this and any initial patchiness in the sampled water was sieving could also remove some chain-forming integrated. diatoms, if present, but since initial inspections After filling the cylinders with 100 1 of seawater, revealed that the phytoplankton was dominated by samples were taken from all the cylinders for immedi- small microflagellates, we considered that removal of ate analyses of nutrients (NOs, NH,, PO, and Si(OH),), extraneous grazers from the containers was more chlorophyll a (chl a), primary production, and particu- important than the error caused by possibly removing late C, N and P. The phytoplankton communities were some chain-forming diatoms. The containers were exposed to different N:P:Si regimes (Table 1). Based on lifted from the ship and placed on a truck for immedi- the concentrations found for the different nutrients in ate transport to the experimental site (this procedure the collected water, additions of PO,, NO, and Si(OH)4 lasted for about 1 h and there was no significant tem- were made to the different cylinders in order to obtain perature change in the water). The enclosures con- different nutrient ratios in the May 1994, June 1993, sisted of 100 1 white polyethylene cylinders with lids, and August 1994 experiments. In the May 1993 exper- and the cylinders were immersed in a plastic swim- iment a surplus of macro- and micronutrients was ming pool supplied with running water in order to added (see Carlsson & Grankli 1999, Turner et al. keep the temperature inside the cylinders close to the 1999). Each treatment was performed in triplicate temperature of the surrounding sea (15 to 18°C in cylinders. Iron, EDTA and trace metals (Cu, Zn, CO, May 1994, 21 to 24°C in June 1993, and 24 to 31°C in Mn and MO) were added at concentrations corre- August 1994). At the experimental site, the water was sponding to 1/20 of the culture medium 'f' (Guillard & transferred from the tanks to the cylinders by siphon- Ryther 1962), in order to prevent any limitation of ing using a 40 mm diameter tube. Water was filled phytoplankton growth by these substances. For the into each cylinder for 20 S at a time, continuously same reason, vitamins (B12, biotin and thiamin) were alternating between the cylinders. In this way the added according to Schone & Schone (1982) (as 1/20 of cylinders were filled with equal amounts of water the original description).

Table 1. Nutrient concentrations in the initial water used in the experiments and the different concentrations reached in the experimental cylinders after nutrient additions. The nutrient additions were either made as 1 initial addition (Expt 1) or as daily additions (Expts 2 and 3). In Expts 2 and 3 the limiting nutrient was added daily while the other nutrients that were in excess were added only when their concentrations were significantly reduced. In the N-deficient Expt 1 treatments, no addition of nitrate was made until the concentration of inorganic nitrogen (nitrate + ammonium) had decreased to 1 pM

Treatment P04 (PM) No3 (PM) SiOz (PM) NH4 (PM)

June 1993 Expt 1 Initial concentrations Mean 0.21 1.03 4.23 1.02 SD 0.01 0.52 0.08 0.09 P-deficient 1 4 0 4 0 N-deficient 2 20 4 0 Si-deficient 2 40 20 Nutrient sufficient 2 40 40 May 1994 Expt 2 Initial concentrations Mean 0.12 3.33 1.39 SD 0 0.12 0.03 P-deficient 0.2 0.2 20 20 P-deficient 0.8 0.8 20 20 N-deficient 1 1.3 1 20 N-deficient 4 1.3 4 20 Si-deficient 1 1.3 20 1 Si-deficient 4 1.3 20 4 Nutrient sufficient 5 0.3 5 5 Nutrient sufficient 20 1.3 20 20 August 1994 Expt 3 Initial concentrations Mean 0.14 1.21 0.72 SD 0 0.20 0.08 P-deficient 0.2 0.2 20 20 N-deficient 1 1.3 1 20 Si-deficient 1 1.3 20 1 Nutrient sufficient 20 1.3 20 20 4 0 Aquat Microb Ecol 18: 37-54, 1999

June 1993 Phosphorus deficient Nitrogen deficient Silicon deficient Nutrient sufficient

Fig. 2.Inorganic nutrient concentrations In the cylinders during the June 1993 expenment (each data po~nt is the mean of 3 replicates, error bars are +l SD; where no error bars are vislble, they are so small that they are hidden Days Days Days Days by the data polnts)

In the 1994 experiments, the water in the cylinders Turner fluorometer model 10 AU after extraction in was filtered once more through a 100 pm nylon net dark for at least 12 h with 95% ethanol according to after 3 d. This was done in order to remove the cope- the method of Jespersen & Christoffersen (1987).Fluo- pods that had developed from nauplii and eggs that rescence units were transformed to pg 1-' of chl a after passed through the initial sieving. This second filtra- spectrophotometric analysis of extracted chl a using tion procedure was designed to prevent a copepod the initial water used in the experiments. Primary pro- population from developing in the cylinders during the duction was measured as I4C-uptake using the method sampling period (as had occurred in the 1993 experi- of Ertebjerg-Nielsen & Bresta (1984). Two pCi of ments). The pool was covered with a black plastic net radioactive NaHI4CO3were added to 25 m1 glass flasks (mesh size of approximately 0.5 cm) that, together with contaming sample water and incubated in the center of the white plastic lids, decreased the light intensity by each cylinder at 0.5 m depth for 2 h (usually between 65 to 70 %. This was expected to prevent photoinhibi- 10:OO and 12:00 h). One light bottle was incubated for tion and excessive solar heating of the phytoplankton. each cylinder and 1 dark bottle was incubated for each Samples were taken daily for analyses of nutrients treatment. Primary production was calculated as pg C (NO3,NH,, PO, and Si(OH),),chl a concentrations, and 1-1 h-~(dark bottle values were subtracted from light for identification and counting of phytoplankton cells bottle values). (see Carlsson & Graneli 1999). Light was measured with a QSL-100 spherical quan- Primary production was measured, and samples tameter (Biospherical Instruments Inc., San Diego). were filtered onto precombusted Gelman A/E filters Light intensity, measured as photosynthetically active (450°C, 2 h) for analyses of particulate C, N and P was radiation (PAR),varied between 80 and 1200 pm01 m-2 also performed at least every third day. In the May S-' in the center of the cylinders during the incuba- 1994 experiment, samples were also taken every third tion~,depending on the time of day and the degree of day for polysaccharide analyses by the methods of cloudiness Analyses of particulate C (POC) and N Dawson & Liebezeit (1981). (PON) were done using a Fisons CHN analyser model Nutrients were analyzed manually following meth- 1108 and the analyses of particulate P (POP) followed ods of Valderrama (1995). Chl a was measured in a the method of Solorzano & Sharp (1980). Graneli et al.: Effects of nutrients and zooplankton grazing on phytoplankton. I 41

RESULTS were supplied at 2 levels, 1 low and 1 higher. In the P- deficient treatments (0.2 and 0.8 pm01 1-' PO, d-') Nutrients phosphate was probably limiting phytoplankton growth during Days 6, 7 and 8 in both treatments. In the June 1993 experiment, nutrient declines re- Phosphate concentration also became very low and flected heavy phytoplankton utilization. Phosphate de- probably limited phytoplankton growth in the Si-defi- creased to low values (~0.25pm01 l-') from Days 4 to 6 cient cylinders where Si was added to reach a concen- in the P-deficient treatment (Fig. 2). For nitrate the con- tration of 4 pm01 I-' d-' (Days 6 and 7) and in the centrations became low (

May 1994 Phosphorus deficient Nitrogen deficient Silicon deficient Nutrient sufficient

SI deficient l pM Nulrisnl sulticent 20 pM

P C

l234567891011

Fig. 3. Inorgan~c nutrient concentrations in the cylinders during the May 1994 experiment (each data point is the mean of 3 replicates; error bars are &lSD; where no error bars are visible, they are so small that they are hidden by the data points) Days Days Days Days 4 2 Aquat M~crobEcol 18: 37-54, 1999

August 1994

Phosphorus deficient Nitrogen deficient Silicon deficient Nutrient sufficient

W 2n 2 r:: iL1&0 0 0 0 01234567891011 01234567891011 01234567891011 ,_p;'-,, 01234567891011

Fig. 4. Inorganic nutrient concentrations in the cylinders dunng the August 05- 05 1994 experiment (each datapornt is the mean of 3 replicates; error bars are *l SD; where no error bars are o m visible, they are so .- small that they are 0 01234567891011 01234567891011 01234567891011 01234567891011 h'ddenb~thedata Days Days Days Days polnts)

concentrations decreased in all treatments and even In the May 1994 experiment, the 2 levels of N defi- to the detection limit of the analysis method at Day 6 ciency (1 and 4 pm01 1-' NO3 added daily) produced the in the N-deficient treatments. However, in all the lowest maximum values of all the treatments, 9.1 and treatments ammonium increased from Day 7 until the 22.1 pg 1-' of chl a, respectively (Fig. 5). The nutrient- end of the experiment. sufficient treatments supplied with the low addition of In the August 1994 experiment (Fig. 4) phosphate nutrients (N:P:Si = 5:0.3:5) produced a maximum value and nitrate were probably limiting phytoplankton slightly higher than that produced by the 4 pm01 1-' N- growth in their respective P- or N-deficient treatments deficient treatment (25.5pg I-'). The treatments where from Days 5 to 11. Only by the last 2 days did Si reach the phosphate and Si were added at low levels (additions same low concentrations found in the May 1994 experi- of 0.2 p011-' P and 1 pm01 1-' Si d-', respectively) pro- ment, i.e. about 0.55 pm01 1-'. In this experiment ammo- duced higher concentrations of chl a (44.6 + 3.3 and nium was reduced to very low or zero values in the N- 43.2 + 1.7 pg I-', respectively). The hlghest chl a con- deficient treatment, while in the other treatments the centrations were found in the nutrient-sufficient treat- ammonium concentrations increased from Day 7 until ment with a higher addition of nutrients (N:P:Si = the end of the experiment, as in the other experiments. 20:1.3:20), and in the P-deficient treatment supplied with 0.8 pm01 1-' P d-': 121.4 k 3.9 and 98.1 t 6.9 pg I-', respectively. Chl a In the August 1994 experiment, N deficiency (1 pm01 1-' No3 added daily) again caused the lowest chl a con- In the June 1993 experiment, chl a concentrations in centrations found in all the treatments (11.1 + 1.5 pg the nutrient-sufficient and the Si-deficient cylinders I-') (Fig. 5). The treatment with the second lowest chl a increased to maximum values of 45.6 + 7.9 (mean +SD) concentration was the Si-deficient treatment (1 p01 and 41.7 * 8.2 pg 1-', respectively (Fig. 5). In the P- and 1-' Si added daily), where maximum chl a values were N-deficient cylinders, somewhat lower chl a concen- 3 times higher than the maximum found in the N-defi- trations of 32.8 + 6.4 and 32.4 + 1.1, respectively, were cient treatment (39.0 ? 1.4 pg I-'). The highest chl a found as maximum values on Day 6 of the experiment. concentrations were found in the P-deficient and nutri- Graneli et al.: Effects of nutrients and zooplankton grazing on phytoplankton. I 43

Phosphorus deficient

U P deficienl 0.2 pt.4 I5O1 1--+- p deficient 0.8 , I'"1

Nitrogen deficient

Silicon deficient

--O-- SI definent l pM + Sl deficient 4 fl

- Nutrient sufficient

(+ Nulrient sunlclenl 5 PM I + Nutrient sufl~c~enr20 jIM

Fig. 5. Chlorophyll a concentrations in the cylinders during the June 1993, May 1994 and August 1994 experiments (each data point is the mean of 3 replicates; error bars are +l SD; Days Days Days where no error bars are visible, they are so small that they are hidden by June 1993 May 1994 August 1994 the data points) ent-sufficient treatments (56.3 k 5.1 pg 1-' and 68.5 + Growth rates 6.0, respectively). Based on increases in the phytoplankton cell C (see Carlsson & Graneli 1999) and chl a, the nutrient-suffi- Primary production cient treatments induced the highest rates of cell divi- sion, followed by the treatments with Si- and P-defi- The primary production rates followed the same cient conditions. The lowest growth rates were found trends as the chl a in the different nutrient treatments when N was the deficient nutrient (Table 2). in all experiments, i.e. the lowest C uptake was found in the N- followed by the P- and Si-deficient cylinders, and the highest values were measured in the nutrient Influence of nutrient ratios and concentrations on sufficient cylinders (Fig. 6). When the inorganic nutri- phytoplankton species composition ents became deficient in the June 1993 experiment, primary production started to decrease 1 d before chl a The general trend from these experiments was that concentrations dropped. small flagellates and diatoms from a few genera 44 Aquat Microb Ecol18: 37-54.1999

Phosphorus deficient

& P deltcient 0.8 pM "1

Nitrogen deficient

I& N delicient 1pM I & N delicient 4 pM "1

Silicon deficient

Nutrient sufficient

& Nutrient sunlclenl 20 PM Fig. 6. Primary production in the cylinders during the June 1993, :lJ,400 l!]/ experimentsMay 1994 and(each August data point 1994 is

, 200 ::LAr,, the mean of 3 repl~cates; error 0 bars are +l SD; where no error 0 123456 7891011 l23 4 5 6 7891011 1234567891011 bars are visible, they are so Days Days Days small that they are hidden by June 1993 May 1994 August 1994 the data points)

Table 2. Growth rates (divisions d-l) based on increase in cell carbon and chlorophyll content for the phytoplankton com- were the most important phytoplankton in the north- munities in the exponential growth phase in the May 1994 ern Adriatic Sea, whereas dinoflagellates were a mesocosm experiment minor component of the phytoplankton in these waters during late spring and summer months. Some Carbon Chl a diatom species were favored under P-deficient conditions, while N deficiency caused some other diatom Nutrient sufficient 1.25 1.70 types to dominate (see Carlsson & Graneli 1999). SI-deficient (1.0 pM Si d-'1 0.68 1 02 Small flagellates were favored by either S1 deficiency (4.0 pM Si d-l) 0.73 1.24 P-deficient (0.2 pM P d-l) 0.68 0 33 (because these flagellates do not require Si as the (0.8 pM P d ') 0.69 0 99 dlatoms do) or by N deficiency. A surplus of N, P and N-deficient (1.0 pM P d.') 0.53 0 0 Si favored the development of some fast-growing (4.0 PM P d-l) 0.58 0.43 diatoms. Graneli et al.. Effects of nutrients and zooplankton grazing on phytoplankton. I 45

May 1994 Phosphorus deficient Phosphorus deficient P dellcient 0.2 pM + P deficient 0.8 pM 1.51

Nitrogen deficient Nitrogen deficient & N deficient l rM & N del~c~enl4 pM

Fig. 7. Total polysaccharides (as pM glucose- amlne equivalents) on Day 10 of the May 1994 experiment (histograms are means of 3 repli- 0- 0 1234S67891011 cates; error bars are k 1 SD) 0 'rL Silicon deficient Silicon deficient Polysaccharide production ca Si dellcient 1 PM m * 3 & SI dellclenl 4 FM One of the most important results found in our enclosure experiments was on polysaccharide production during the May 1994 experiment. The highest amount of total polysaccharides produced was under nutrient- sufficient and P-deficient conditions (Fig. ?). However, N deficiency induced the phyto- Nutrient sufficient Nutrient sufficient plankton in our cylinders to produce more polysaccharides per cell C or chl a (Fig. 8). * Nutrient su~lclenl20 pM

Phytoplankton chemical composition

The phytoplankton chemical composition , , changed from the beginning to the end of the 0- 1234567891011 1234567891011K, experiments. These changes were more ac- Days Days centuated between the different nutrient treatments than between experiments. The May 1994 May 1994 POC. PON and POP increased steadily Fig. 8. Total po1ysaccharides:POC (weight:weight) and total polysaccha- toward the end of the experiments in the rides:chl a (weight:weight) in the May 1994 experiment (each data point nutrient-sufficient conditions in the is the mean of 3 replicates; error bars are rl SD;where no error bars are treatments where the limiting nutrient was visible, they are so small that they are hidden by the data points) added at the higher levels to the tanks (Figs. 9, 10 & 11). The exception was when N was added gan 1995) and our experiments support this in that P daily at the lowest concentration (1 poll-' N d-') where additions in low (0.2 pm01 1-' P d-l) or high (0.8 to POC and PON were found at the lowest concentrations. 1.3 pm01 1-' P d-l) amounts increased phytoplankton biomass. However, this was only possible when N was kept at high concentrations (above 10 pm01 1-l). When DISCUSSION N was added in low concentrations (1 pm01 1-l d-l), the resulting phytoplankton biomass was much lower than The northern Adriatic Sea is considered to be P lim- when P was added at the low concentrations found in ited (Pojed & Kveder 1977, Thingstad & Rassoulzade- situ during the sampling for the first experiment. In 4 6 Aquat Microb Ecol 18: 37-54, 1999

Phosphorus deficient

Nitrogen deficient + N deficient 1 &M

200

0 1234561891011 01234567891011 01234567891011

0 Silicon deficient

Nutrient sufficient

& NuIrIenl Sufficient 20 WM Fig. 9. Particulate organic carbon (POC) during the June !!l/,300 1993, May 1994 and August 200 1994 experiments (each data 100 point is the mean of 3 replicates; 0 , , i/J, 0 123456789l011 012'34567891011 Ol 234567891011 error bars are *l SD; where no Days Days ;/pDays error bars are visible, they are so small that they are hidden by June 1993 May 1994 August 1994 the data points)

terms of environmental policy management, it was ited. The latter might have been the case as diatoms only the 'low' levels of N that could keep the phyto- dominated in almost all treatments in all experiments. plankton biomass under control. The concentrations of Si in the northern Adriatic Sea The phytoplankton was never able to reduce Si con- are between 4.7 * 1.5 and 2.6 + 0.9 pm01 1-' in surface centrations below 0.55 pm01 I-', a concentration that and intermediate-depth layers, respectively, during was measured from Days 5 to 10 for all treatments with summer periods (Franco & Michelato 1992). This Si-deficiency, and in the N:P:Si 5:0.3:5 treatment. Our makes it unlikely that Si would ever limit diatom measurements of concentrations in standards were growth in these waters. performed on artificial seawater, without any problem Ammonium increased toward the end of the experi- measuring Si concentrations below these levels. Thus, ments, except for the N-deficient treatment in the Au- either the diatoms in the Adriatic were unable to utilize gust 1994 experiment, when ammonium concentrations Si at these concentrations or the cells were not Si lim- were low throughout the entire experiment. This in- Graneli et al.: Effects of nutrients and zooplankton grazing on phytoplankton. I 4 7

Phosphorus deficient

U P deficient 0 2 pM

100 P

Nitrogen deficient

Silicon deficient

Nutrient sufficient IU Nutrient sunlclent 5 NM I

Fig. 10. Particulate organic nitrogen (PON) during the June 1993, May 1994 and August 1994 experiments (each data point is the mean of 3 replicates; error bars are +l SD; where no error bars are visible, they Days Days Days are so small that they are hidden by the data points) June 1993 May 1994 August 1994

crease in ammonium was possibly a result of suboptimal year there are high concentrations of both inorganic N conditions for some of the phytoplankton under different and phosphate occurring in the surface water (see nutrient-deficient conditions, causing higher hetero- >l0 yr of monitoring data in the reports of Regione trophlc activity by bacteria and heterotrophic flagellates. Emilia Romagna, Franco & Michelato 1992). Thus, the The ratio for the inorganic N to P compounds in the normal situation is that neither phosphate nor inor- Adriatic Sea is usually higher than the Redfield ratio, ganic N are totally depleted. The conclusion from this suggesting that P could limit the production of phyto- is that the phytoplankton usually will not deplete the plankton biomass (Chiaudani et al. 1980b). Bioassay water of either inorganic N or P, even if a high phyto- experiments have confirmed that P can limit phytoplankton biomass is present. The notion that P should plankton biomass (Chiaudani & Vighi 1982, Mingazz- limit the phytoplankton production is therefore only ini et al. 1992).However, extensive field data from the true during the periods of the year when the phyto- northern Adriatic shows that at almost all times of the plankton have used all the available P. 4 8 Aquat Microb Ecol 18: 37-54, 1999

Phosphorus deficient

+ P dellcienl 0.8 pM

Nitrogen deficient

N delicienl 4 pM 12 +

Silicon deficient

Nutrient sufficient

+ Nulrienl sulficisnl 20 PM 10 Fig. 11 Particulate organic phosphorus (POP) during the June 1993, May 1994 and August 1994 experiments (each data point 1s the mean of 3 replicates; error bars are *l SD; Days Days Days where no error bars are visible. they are so small that they are June 1993 May 1994 August 1994 hidden by the data points)

Why then are neither phosphate nor inorganic N minor factor limiting phytoplankton depletion of the fully utilized during most of the growth season? The inorganic nutrients in the northern Adriatic Sea. Al- explanation for this may be that the phytoplankton ternatively, the filtration activity of the benthos is cells are removed from the water column faster than potentially substantial, and it has been shown that all they grow, either by zooplankton grazing or by the fil- phytoplankton can be removed from a 10 m deep, tration activity from the benthos. The grazing expen- well-mixed water column by the benthic suspension ments in our study (see Turner et al. 1999) showed that feeders in 1 to 3 d in Scandinavian waters (Loo & even 10 times higher than normal mesozooplankton Rosenberg 1989). Perhaps in the shallow northern part abundance could not curtail phytoplankton growth in of the Adriatic Sea, the high biomass of benthic sus- our enclosures, and that the phytoplankton ultimately pension feeders may have the capacity to remove such have the potential to use all inorganic nutrients pre- large quantities of phytoplankton that no real limita- sent. Therefore it seems that zooplankton grazing is a tion of either N or P normally occurs. Graneli et al.: Effects of nutrients and zooplankton grazing on phytoplankton. I 49

Phosphorus deficient

Nitrogen deficient

Silicon deficient

Nutrient sufficient

Nutrlenl SufficientS PM

Nulrlenl sufficienl 20 IIM

Fig. 12. P0C:chl a ratios during the June 1993. May 1994 and August loo 1994 experiments (each data point is , the mean of 3 replicates; error bars 01234567891011 1234567891011 01214567891011 are *l SD; where no error bars are Days Days Days visible, they are so small that they are hidden by the data points) June 1993 May 1994 August 1994

In our experiments, diatom growth was likely only (Paasche 1980). The conclusion is that the diatoms marginally affected by low Si concentrations in the Si- could grow quite well in our experiments in spite of low deficient treatments. This can be seen in the P0C:chl a Si concentrations. In the June 1993 experiment, how- ratios (Fig, 12), which were always below 100, indicat- ever, diatoms were not so numerous in the initial com- ing healthy, growing cells (Goldman 1980), and in the munity, and increased in treatments with high Si addi- high growth rates, based on increases in chl a, which tion. However, the small autotrophic flagellates kept were always >l d-' (Table 2).In the Si-deficient treat- their dominance in the Si-deficient treatment, indicat- ments, diatoms used the Si that was added each day ing that lack of Si could restrain the growth of diatoms such that silicate concentrations were reduced to 0.5 to in this experiment (see Carlsson & Graneli 1999). 0.6 pm01 1-' the day after additions. This is in the range Different diatom species have been shown to have of, or slightly higher than, reported half saturation con- different demands for Si compared to P (Si:P = 96:l to stants for various marine diatoms of 0.02 to 0.5 pm01 1-' 1:1, Kilham & Hecky 1988). If we assume that the N:P 50 Aquat Microb Ecol 18: 37-54, 1999

demand is 16:l (Redfield ratio), then diatoms with a to be the thick cell wall of cellulose (Sakshaug et al. low Si:P demand (around 1) would need very small 1984).In our cylinders diatoms made up the bulk of the amounts of Si compared to N. biomass in most experiments. Thus, the variations in In the P-deficient treatment with the hlghest daily P ad- our ratios were not as large as those found by Duarte dition (0.8 pm01 l-' d-'), the phytoplankton never became (1992). The variations in C:N:P ratios in our tanks P-deficient (P0C:chl a was always below 100) (Fig. 12). mostly depended upon whether the phytoplankton were nutrient limited or not, and the values differed maximally by about 2-fold. For instance, PON: POC Growth rates and phytoplankton and POP : POC ratios either remained at the same levels chemical composition or decreased in the nutrient-deficient treatments. This occurred in the N- and P-deficient cylinders for the May Nutrient deficiency or sufficiency in phytoplankton 1994 and August 1994 experiments. This suggests that is not only defined by the nutrient concentrations and the phytoplankton in these treatments were growing ratios in the medium where the phytoplankton are suboptimally due to N or P limitation, and accumulating growing, but also in terms of the intracellular chemical excess C from photosynthesis at the same time. composition. The latter gives a better indication nu- The P0N:chl a and P0P:chl a ratios (Figs. 13 & 14) trient limitation (Droop 1974). Many studies using decreased from high initial values to low ones during cultures have shown that it is difficult to relate phyto- the middle of the experiments. These ratios increased plankton growth rates to external nutrient concentra- again most markedly in the N- and P-deficient treat- tions, and that growth rates are related instead to the ments, although less markedly in the latter. Silicon internal nutrient conditions (Hecky & Kilham 1988 and deficiency did not provoke such strong response in the references therein). phytoplankton cells. The reason seems to be that, even Different chemical compositions (C:N:P) between if the diatoms made up the bulk of the phytoplankton and within different algal groups may explain differ- biomass in all the containers, they could grow at low Si ences in composition of particulate material between concentrations without suffering from severe Si defi- different treatments in our experiments (Healy & ciency. This is also supported by the fact that the Hendzel 1979, Hecky & I(llham 1988). Also, luxury POP: POC or PON: POC ratios in the Si-deficient treat- consumption of P may give very low N:P ratios (Mack- ments did not increase markedly at the end of the ereth 1953, Reynolds 1984). experiments. The decrease seen in the P0P:chl a and Some phytoplankton species appear to be uptake spe- P0N:chl a ratios indicates that initially there was a cialists, whereas others are storage specialists (Tilman et pool of PON (between 1.5 to 2.9 pm01 I-') and POP al. 1982, Sornmer 1986). The pulsed addition of nutrients (between 0.3 and 0.8 pm01 1-') which probably was not in our experiment should have been beneficial to the in the phytoplankton but rather in detritus. species adapted to rapidly take up nutrient pulses. The POC: chl a ratios were also higher initially in all In our experiments, as expected, the P0C:chl a ratios treatments, indicating that a substantial pool of detri- were very high in the N-deficient cylinders (lowest = tus or other particulate material unrelated to the 80:1, weight basis in August 1994). This indicates that phytoplankton was present. However, after a few days the cells were N-starved and unable to build up chl a when the phytoplankton biomass increased, these due to the lack of N. Goldman (1980) proposed that ratios reached levels considered to be normal for non- such low values could not support growth rates (p) nutrient-limited phytoplankton cells (between 40 and above 0.1. Indeed p was zero in the May, 1994 N-defi- 60 on a weight basis). Toward the end of the experi- cient cylinders where 1 pm01 1-' N was added daily, but ments these ratios increased again. Inorganic nutrient 0.43 in the N-deficient treatment where 4 pm01 1-' N concentrations in the tanks during the experiments in was added daily, although the P0C:chl a ratio was May 1994 and August 1994 suggested that the phyto- above 100 in both treatments in August 1994. plankton cells were nutrient limitated in the P- and N- The ratios of C:N:P showed a different pattern from deficient cylinders. However, in the June 1993 experi- their concentrations in the particulate material. Duarte ment the POC: chl a ratios did not increase at all. This (1992), comparing intracellular concentrations and ra- experiment was different from the others, however, in tios for C:N:P for 96 phytoplankton species from the lit- that after the fifth day we did not mix the water in the erature, found that there was a broad range of variation cylinders before sampling. Samples were taken both for the ratios of these elements. Carbon was the ele- from the surface water and sedimented material in an ment showing the highest variability for the different attempt to see if mucilage was being produced by sed- groups. Diatoms had the lowest concentrations while irnenting diatoms during stable conditions. We did not the highest were found for dinoflagellates and green al- observe any mucilage formation by the sedimenting gae. The explanation seems, at least for dinoflagellates, diatoms, however, in any of the treatments. Graneli et al.: Effects of nutrients and zooplankton grazing on phytoplankton. I 5 1

Phosphorus deficient + P delicienl 0.2 PM & P delicienl 0.8 pM

Nitrogen deficient

+ N delicienl 4 W

L .-m Silicon deficient a, YS

Nutrient sufficient

Nulnent sullicienl 5 pM

Fig. 13. P0N:chl a ratios during the June 1993, May 1994 and 20 August 1994 experiments (each data point is the mean of 3 repli- 0 123 4567 891011 1234561891011 01234567891011 cates; error bars are i1SD; where - Days Days Days no error bars are visible, they are so small that they are hidden by the data points) June 1993 May 1994 August 1994

Polysaccharide production pelagic diatoms, and they were able to produce high quantities of polysaccharides. We do not know The mucilage formation in the northern Adriatic Sea whether benthic diatoms would produce the same has long been believed to be due to exudates of ben- amount or even more, since benthic diatoms did not thc diatoms (Forti 1906, Zanon 1931). Conversely, comprise a substantial part of the microalgae biomass. Degobbis (1995) collected information from several The higher amounts of polysaccharides produced independent sources and came to the conclusion that per unit of cell C were under N-deficiency. This agrees mucilage is a result of polysaccharide production by with reports from elsewhere which show that exudate pelagic and not benthic diatoms. This theory fits well production is stimulated when the algae are first sub- with our results. The phytoplankton in our cylinders jected to high N:P ratios, and thereafter are subjected during the May 1994 experiment was dominated by to low nutrient concentrations (Myklestad 1977, Monti 52 Aquat Microb Ecol18: 37-54, 1999

Phosphorus deficient

--t P delicienr 0.8 fl

Nitrogen deficient

N dellclenl 1pM

Silicon deficient

+ SI dellclent 4 pM

Nutrient sufficient

I--t Nutrient sunlcient 20 pM I & Nutrienl sulltcienl 5 pM

Fig 14 P0P:chl a ratios during the June 1993, May 1994 and August 1994 experiments (each data point is the

01234567891011 1234567891011 OI~~~~~~~~~~~~meanof3replicates;errorbarsare *l SD; where no error bars are visible, Days Days Days they are so small that they are hidden June 1993 May 1994 August 1994 by the data points) et al. 1995). Thus, as Degobbis (1995) suggested, the area. This is despite the fact that P is in 'limiting' high production of polysaccharides seems to be related amounts in relation to the 16:l optimal Redfield ratio for to runoff pulses (mostly from the PO river), followed by N and P thought to be beneficial for phytoplankton a period with nutrient depletion. growth. Since a low daily addition of P produced approximately double the biornass of phytoplankton compared to an equivalent Redfield N addition, only when Conclusions a high concentration of inorganic N was present, we suggest that a reduction of N discharges to the northern The input of inorganic nutrients (N and P) to coastal Adriatic Sea would result in a reduced eutrophication areas of the northern Adriatic Sea appears to be suffi- situation in the area. The anomalous production of cient to produce the large algal biomass responsible for phytoplankton induced by P addition compared to N the large production of polysacchandes seen in the addition may well be caused by the fast recycling of P. Graneli et al.. Effects of nutrients and zooplankton grazing on phytoplankton. I 53

Acknowledgements. Funding was from the EEC (European Forti A (1906) Alcune osservazioni sul 'mare sporco' ed in par- Economic Community), contract no. EV5V-CT92-0215, ticolare sul fenomeno avvenuto nel 1905. Nouvo G Bot ltal MARE Project, and the funds were made available through (NS) 13:l-56 the Swedish Environmental Protection Agency. We appreci- Franco P, Michelato A (1992) Northern Adriatic Sea: oceanog- ate the assistance of Simonetta della Libera, Rafaele Scenati, raphy of the basin proper and of the western coastal zone. Giovanni Costamagna, and Milena Bruno (Istituto Superiore In: Vollenweider RA, Marchetti R, Viviani R (eds) Marine di Sanita, Rome), Ludovica Diliberto (Centro di hcerca e For- coastal eutrophication. Elsevier, New York, p 36-52 mazione peril Control10 dei Sistemi Idrici, Sardegna) and the Gilmartin M, Revelante N (1980) Nutr~entinput and the sum- staff of the Istituto Zooprofilatico Delle Umbria e de La mer nanoplankton bloom in the Northern Adriatic Sea. Marche, Fano. Farooq Azam provided valuable criticism of PSZN I: Mar Ecol 1:169- 180 the manuscript. Goldrnan JC (1980) Physiological processes, nutrient availabil- ity and the concept of relative growth rate in marine phyto- LITERATURE CITED plankton ecology. Brookhaven Symp Biol31:179-194 Graneli E, Wallstrom K, Larsson U, Graneli W. Ehngren R Artebjerg-Nielsen G, Bresta AM (eds) (1984) The Baltic (1990) Nutrient limtation of primary production in the marine biologists. Guidelines for the measurement of Baltic Sea area. Ambio 19:142-151 phytoplankton primary production. BMB Pub1 no. 1, 2nd Gulllard RRL, Ryther JH (1962) Stu&es of marine planktonic edn. Marine Pollution Laboratory, Charlottenlund &atoms I Cyclotella nand Hustedt, and Detonula confer- Berland BR, Bonin DJ, Maestrini SY (1980) Azote ou phos- vacea (Cleve) Gran. Can J Microbiol 8:229-239 phore? Considerations sur le 'paradoxe nutritionnel' de la Healy FP, Hendzel LL (1979) Indcators of phosphorus and mer mediterranee. Oceanol Acta 3:135-142 nitrogen deficiency in five algae in culture. J Fish Res Boni L, Milandri A, Poletti R, Pompei M (1993) DSP-cases Board Can 36:1364-1369 along the Emiglia-Romagna coast (Northwestern Adriatic Hecky RE, Kilham P (1988) Nutrient limitation of phytoplank- Sea). In: Smayda TJ, Shirnizu Y (eds) Toxic phytoplankton ton in freshwater and marine environments: a review of blooms in the sea. Elsevier, Amsterdam, p 475-481 recent evidence on the effects of enrichment. Limnol Caraco N, Cole J, Likens GE (1990) A comparison of phos- Oceanogr 33:796-822 phorus immobihzation in sedirnents of freshwater and Howarth RW (1988) Nutrient limitation of net primary pro- coastal marine systems. J Biochem 9:277-290 duction in marine ecosystems. Ann Rev Ecol 19:89-110 Carlsson P, Graneli E (1999) Effects of N:P:Si ratios and zoo- Jespersen AM, Christoffersen K (1987) Measurements of chl- plankton grazing on phytoplankton communities in the a from phytoplankton using ethanol as extraction solvent. northern Adriatic Sea. I1 Phytoplankton species composi- Arch Hydrobiol 109:445-454 tion. Aquat Microb Ecol 1855-65 Justic D (1987) Long-term eutrophication of the Northern Cati L (1981) Idrografia e idrologia del PO. Pubbli Ufficio Idro- Adriatic Sea. Mar Pollut Bull 18:281-284 grafico del PO, Minister0 Lavori Pubblici, Rome Justic D (1991) A simple oxygen index of trophic state Chiaudani G. Vighi M (1982) Multistep approach to identifi- description. Mar Pollut Bull 22:201-204 cation of limiting nutrients in Northern Adriatic eutro- Kilham P, Hecky RE (1988) Nutrient limitation of phytoplank- phied waters. Water Res 21:219-229 ton in freshwater and marine environments: a review of Chiaudani G, Marchetti R, Vighi M (1980a) Eutrophication in recent evidence on the effects of enrichment. Limnol Emilia-Romagna coastal waters (North Adriatic Sea, Italy): Oceanogr 33:776-795 a case history. Prog Wat Technol 12 185-192 Likens GE (ed) (1972) Nutrients and eutrophication: the limit- Chiaudani G, Gaggino GF, Vighi M (1980b) Nutnent limita- ing-nutrient controversy. Am Soc Limnol Oceanogr Spec tion assessment in Ernilia-Romagna eutrophicated coastal Symp. Allen Press, Inc, Lawrence, KS waters. In: Workshop on pollution of the Mediterranean. Loo LO, Rosenberg R (1989) Bivalve suspension-feeding Proc ICSEM/UNEP. Cagliari, p 383-390 dynamics and benthic-pelagic coupllng in an eutrophi- Dawson R, Liebezeit G (1981) The analytical methods cated marine bay. J Exp Mar Biol Ecol 130:253-276 for the characterisation of organics in seawater. In: Mackereth FJH (1953) Phosphorus utilization by Astenonella Duursma EK Dawson R (eds) Marine organic chemistry. formosa Has. J Exp Bot 4:296-313 Evolution, composition interactions and chemistry of or- Malej A, Mozetic P, Malacic V, Terzic S, Ahel M (1995) Phyto- ganic matter in seawater. Elsevier Oceanography Series plankton responses to freshwater inputs in a small semi- 31:463-465 enclosed gulf (Gulf of Trieste, Adriatic Sea). Mar Ecol Prog Degobbis D (1989) Increased eutrophication of the Northern Ser 120:lll-121 Adriatic Sea. Mar Pollut Bull 20:452-457 Marchetti R (1992) The problems of the Emilia Romagna Degobbis D (1995) Increased eutrophication of the Northern coastal waters: facts and interpretations. In: Vollenweider Adriatic Sea-second act. Mar Pollut Bull 9:452-457 RA, Marchetti R, Viviani R (eds) Marine coastal eutrophi- Degobbis D, Gilmartin M (1990) Nitrogen, phosphorus, and cation. Elsevier, Amsterdam, p 21-33 biogenic silicon budgets for the Northern Adriatic Sea. Mingazzini M, Rinaldi A, Montanari G (1992) Multi-level Oceanol Acta 13:31-45 nutrient enrichment bioassays on Northern Adriatic Droop MR (1974) The nutrient status of algal cells in continu- coastal waters. In: Vollenweider RA. Marchetti R, Viviani ous culture. J Mar Biol Assoc UK 54:825-855 R (eds) Marine coastal eutrophication. Elsevier, Amster- Duarte CM (1992) Nutrient concentration of aquatic plants: dam, p 115-131 patterns across species. Lirnnol Oceanogr 37:882-889 Monti M. Welker C, Dellavalle G, Casaretto L, Fonda-Umani Fonda Umani S, Franco P, Ghlrardelh E, Malej A (1993) Out- S (1995) Mucous aggregates under natural and laboratory line of oceanography and the plankton of the Adriatic Sea. conditions: a review. Sci Total Environ 165:145-154 In: Colombo G, Ferrari I, Ceccherelli VU, Rossi R (eds) Myklestad S (1977) Production of carbohydrates by marine Marine eutrophication and population dynamics. Proc phytoplanktonic diatoms. I1 Influence of the N/P ratio in 25th Eur Mar Biol Symp. Olsen & Olsen, Fredensborg, the growth medium on the assimilation ratio, growth rate, p 347-365 and production of cellular and extracellular carbohydrates 54 Aquat Microb Ecol 18: 37-54, 1999

by Chaetoceros affinis var. Willei (Gran) Hustedt and Smith SV (1984) Phosphorus versus nitrogen limitation in the Skeletonema costatum (Grev.) Cleve. J Exp Mar Biol Ecol marine environment. Limnol Oceanogr 29:1149-1160 29:161-179 Solorzano L, Sharp JH (1980) Determination of total dissolved Nixon SW (1990) Marine eutrophication: a growing interna- phosphorus and particulate phosphorus in natural waters. tional problem. Ambio 19:l-176 Lirnnol Oceanogr 25:754-758 Olsson P, Graneli E, Carlsson P, Abreu P (1992) Structuring Sornmer U (1986) Phytoplankton competition along a gradient of a postspring phytoplankton community by manipula- of dilution rates Oecologia 68:503-506 tion of trophic interactions. J Exp Mar Biol Ecol 158: Stachowitsch M, Fanuko N, Richter M (1990) Mucus aggre- 249-266 gates in the Adriatic Sea: an overview of stages and occur- Oviatt C, Doering P, Nowicki B, Reed L, Cole J, Frithsen J rences. PSZN I: Mar Ecol 11:327-350 (1995) An ecosystem level experiment on nutrient lirnita- Thingstad TF, Rassoulzadegan F (1995) Nutrient limitation, tion in temperate coastal marine environments. Mar Ecol microbial food webs, and 'b~ologicalC-pumps': suggested Prog Ser 116:1?1-179 interactions in a P-limited Mediterranean. Mar Ecol Prog Paasche E (1980) Silicon. In: Morns 1 (ed) The physiological Ser 117:299-306 ecology of phytoplankton. Blackwell, Oxford, p 259-284 Tilman D, Kilham SS, Kdham P (1982) Phytoplankton commu- Piccinetti C, Manfrin G (1969) Osservazioni sulla mortalita di nity ecology: the role of limiting resources. Annu Rev Ecol pesci e di altri organismi verificatasi nel 1969 in Adriatico. Syst 13:349-372 Note Lab Biol Mar Pesca-Fano 3:73-92 Turner JT, Tester PA, Lincoln JA, Carlsson P, Graneli E (1999) Pojed I, Kveder S (1977) Investigation of nutrient limitation of Effects of N:P:Si ratios and zooplankton grazing on phyto- phytoplankton production in the North Adriatic by enrich- plankton communities in the northern Adriatic Sea. 111. ment experiments. Thalassia Jugosl 13:13-24 Zooplankton populations and grazing. Aquat Microb Ecol Reynolds CS (ed) (1984) The ecology of freshwater phyto- 18:67-75 plankton. Cambridge studies in ecology. Cambridge Uni- Valderrama JC (1995) Methods of nutrient analysis. In: Halle- versity Press, Cambridge graeff GM, Anderson DM. Cernbella AD (eds) Manual of Ryther JH, Dunstan WM (1971) Nitrogen, phosphorus and harmful marine rnicroalge. IOC Manuals and Guides No. eutrophication in the coastal marine environment. Science 33, UNESCO Workshop, Paris, p 251-268 171:1008-1013 Vollenweider RA, Rinaldi A, Montanari G (1992) Eutrophica- Sakshaug E, Graneli E, Elbrachter M, Kayser H (1984) Chem- tion, structure and dynamics of a marine coastal system: ical composition and alkallne phosphatase activity of results of ten-year monitonng along the Emilia-Romagna nutrient-saturated and P-deficient cells of four marine coast (Northwest Adriatic Sea). In: Vollenweider RA, Mar- dinoflagellates. J Exp Mar Biol Ecol?7:241-254 chetti R, Viviani R (eds) Marine coastal eutrophication. Schone HK, Schone A (1982) A weakly enriched sea-water Elsevier, Amsterdam, p 63-106 medium for ecological studies on marine plankton algae, Zanon DV (1931) Esame di un campione di 'mare sporco' del some examples of its application. Bot Mar 25:117-122 Golfo di Fiume. Mem Pont Acad Lincei 15:l-33

Editorial responsibility: John Dolan, Submitted: September 22, 1997; Accepted: October 29, 1998 Villefranche-sur-Mer, France Proofs received from author(s): July 12, 1999